Average Error: 14.3 → 0.0
Time: 6.2s
Precision: binary64
\[\left(0 \leq b \land b \leq a\right) \land a \leq 1\]
\[\sqrt{\left|\frac{a \cdot a - b \cdot b}{a \cdot a}\right|} \]
↓
\[\sqrt{\left|\mathsf{fma}\left(b, \frac{1}{a} \cdot \frac{b}{a}, -1\right)\right|} \]
\sqrt{\left|\frac{a \cdot a - b \cdot b}{a \cdot a}\right|}
↓
\sqrt{\left|\mathsf{fma}\left(b, \frac{1}{a} \cdot \frac{b}{a}, -1\right)\right|}

(FPCore (a b)
 :precision binary64
 (sqrt (fabs (/ (- (* a a) (* b b)) (* a a)))))
↓
(FPCore (a b)
 :precision binary64
 (sqrt (fabs (fma b (* (/ 1.0 a) (/ b a)) -1.0))))
double code(double a, double b) {
	return sqrt(fabs(((a * a) - (b * b)) / (a * a)));
}
↓
double code(double a, double b) {
	return sqrt(fabs(fma(b, ((1.0 / a) * (b / a)), -1.0)));
}

Error

Bits error versus a

Bits error versus b

Derivation

  1. Initial program 14.3

    \[\sqrt{\left|\frac{a \cdot a - b \cdot b}{a \cdot a}\right|} \]

  2. Simplified14.3

    \[\leadsto \color{blue}{\sqrt{\left|\mathsf{fma}\left(b, \frac{b}{a \cdot a}, -1\right)\right|}} \]

  3. Applied *-un-lft-identity_binary6414.3

    \[\leadsto \sqrt{\left|\mathsf{fma}\left(b, \frac{\color{blue}{1 \cdot b}}{a \cdot a}, -1\right)\right|} \]

  4. Applied times-frac_binary640.0

    \[\leadsto \sqrt{\left|\mathsf{fma}\left(b, \color{blue}{\frac{1}{a} \cdot \frac{b}{a}}, -1\right)\right|} \]

  5. Final simplification0.0

    \[\leadsto \sqrt{\left|\mathsf{fma}\left(b, \frac{1}{a} \cdot \frac{b}{a}, -1\right)\right|} \]

Reproduce

herbie shell --seed 2022055 
(FPCore (a b)
  :name "Eccentricity of an ellipse"
  :precision binary64
  :pre (and (and (<= 0.0 b) (<= b a)) (<= a 1.0))
  (sqrt (fabs (/ (- (* a a) (* b b)) (* a a)))))